Re: [Opensvcdecoder-support] OSVC unable to decode a video that JSVM can

[Brett S. <bs...@uo...>]

I found my mistake.

I had accidentally types 172 instead of 176 for the resolution of the base
layer. Since this was not a factor of 2, that meant the baseline
requirement of OSVC was broken, which would explain why JSVM worked.

I will test more, but I think it is working now, so please disregard my
question. Sorry for the needless post.

On Thu, Jul 12, 2012 at 3:23 PM, Brett Slote <bs...@uo...> wrote:

> (failed to post due to attachment, removed & replaced with link)
>
>
>
> Hello,
>
> I'm having trouble with a .264 file that JSVM decodes fine that OSVC
> won't. Any help would be greatly appreciated, as always. I hope it is just
> something in my config file not supported by OSVC that can easily be
> avoided.
>
> I have a small h.264 SVC video- foreman0.264 (8 frames of foreman) with 3
> layers and no loss.
>
> Both JSVM and OSVC decode it fine.
>
> I use JSVM's BitStreamExtractorStatic32 to produce a .264 file with only
> the base layer. (foremansBase.264)
>
> I also drop all packets except those of the base layer in my network
> simulation. (decInp.264)
>
> So I have 2 files that should just be the base layer, with no loss
> otherwise.
>
> Both decode fine in JSVM, producing foremansBase.yuv and decInp.yuv
> (terribly quality due to high qp, but valid).
>
> However, OSVC produces corrupt results look like diagonal lines of the
> image (openSVCOut0.yuv). I have verified that it looks the same when I play
> it in MPlayer as included with OSVC, so it is OSVC and not my code for
> feeding OSVC.
>
> All files plus the config files are attached, or at www.slote.ca/temp.zip(650 kb).
>
> Here is the profile and level info from encoding:
>
> profile & level info:
>
> =====================
>
> DQ=  0:  Main @ Level 1.1
>
> DQ= 16:  Scalable High @ Level 1.1
>
> DQ= 32:  Scalable High @ Level 2.1
> Thank you for any help or insight you can offer,
>
> Brett Slote
>
>
>